Patch 594001: PEP 277 - Unicode file name support for Windows NT.

Include/pyerrors.h

@@ -81,16 +81,37 @@ PyAPI_DATA(PyObject *) PyExc_FutureWarning;
 PyAPI_FUNC(int) PyErr_BadArgument(void);
 PyAPI_FUNC(PyObject *) PyErr_NoMemory(void);
 PyAPI_FUNC(PyObject *) PyErr_SetFromErrno(PyObject *);
+PyAPI_FUNC(PyObject *) PyErr_SetFromErrnoWithFilenameObject(
+	PyObject *, PyObject *);
 PyAPI_FUNC(PyObject *) PyErr_SetFromErrnoWithFilename(PyObject *, char *);
+#ifdef Py_WIN_WIDE_FILENAMES
+PyAPI_FUNC(PyObject *) PyErr_SetFromErrnoWithUnicodeFilename(
+	PyObject *, Py_UNICODE *);
+#endif /* Py_WIN_WIDE_FILENAMES */
+
 PyAPI_FUNC(PyObject *) PyErr_Format(PyObject *, const char *, ...)
 			Py_GCC_ATTRIBUTE((format(printf, 2, 3)));
+
 #ifdef MS_WINDOWS
-PyAPI_FUNC(PyObject *) PyErr_SetFromWindowsErrWithFilename(int, const char *);
+PyAPI_FUNC(PyObject *) PyErr_SetFromWindowsErrWithFilenameObject(
+	int, const char *);
+PyAPI_FUNC(PyObject *) PyErr_SetFromWindowsErrWithFilename(
+	int, const char *);
+#ifdef Py_WIN_WIDE_FILENAMES
+PyAPI_FUNC(PyObject *) PyErr_SetFromWindowsErrWithUnicodeFilename(
+	int, const Py_UNICODE *);
+#endif /* Py_WIN_WIDE_FILENAMES */
 PyAPI_FUNC(PyObject *) PyErr_SetFromWindowsErr(int);
+PyAPI_FUNC(PyObject *) PyErr_SetExcFromWindowsErrWithFilenameObject(
+	PyObject *,int, PyObject *);
 PyAPI_FUNC(PyObject *) PyErr_SetExcFromWindowsErrWithFilename(
 	PyObject *,int, const char *);
+#ifdef Py_WIN_WIDE_FILENAMES
+PyAPI_FUNC(PyObject *) PyErr_SetExcFromWindowsErrWithUnicodeFilename(
+	PyObject *,int, const Py_UNICODE *);
+#endif /* Py_WIN_WIDE_FILENAMES */
 PyAPI_FUNC(PyObject *) PyErr_SetExcFromWindowsErr(PyObject *, int);
-#endif
+#endif /* MS_WINDOWS */
 
 /* Export the old function so that the existing API remains available: */
 PyAPI_FUNC(void) PyErr_BadInternalCall(void);

Lib/test/test_unicode_file.py

@@ -40,6 +40,10 @@ os.chmod(TESTFN_ENCODED, 0777)
 os.chmod(TESTFN_UNICODE, 0777)
 
 # Test rename
+try:
+    os.unlink(TESTFN_ENCODED + ".new")
+except os.error:
+    pass
 os.rename(TESTFN_ENCODED, TESTFN_ENCODED + ".new")
 os.rename(TESTFN_UNICODE+".new", TESTFN_ENCODED)
 

Objects/fileobject.c

@@ -15,6 +15,12 @@
 #include <windows.h>
 #endif
 
+#ifdef _MSC_VER
+/* Need GetVersion to see if on NT so safe to use _wfopen */
+#define WIN32_LEAN_AND_MEAN
+#include <windows.h>
+#endif /* _MSC_VER */
+
 #ifdef macintosh
 #ifdef USE_GUSI
 #define HAVE_FTRUNCATE
@@ -102,7 +108,7 @@ dircheck(PyFileObject* f)
 
 static PyObject *
 fill_file_fields(PyFileObject *f, FILE *fp, char *name, char *mode,
-		 int (*close)(FILE *))
+		 int (*close)(FILE *), PyObject *wname)
 {
 	assert(f != NULL);
 	assert(PyFile_Check(f));
@@ -110,6 +116,9 @@ fill_file_fields(PyFileObject *f, FILE *fp, char *name, char *mode,
 
 	Py_DECREF(f->f_name);
 	Py_DECREF(f->f_mode);
+	if (wname)
+		f->f_name = PyUnicode_FromObject(wname);
+	else
 		f->f_name = PyString_FromString(name);
 	f->f_mode = PyString_FromString(mode);
 
@@ -135,7 +144,12 @@ open_the_file(PyFileObject *f, char *name, char *mode)
 {
 	assert(f != NULL);
 	assert(PyFile_Check(f));
+#ifdef MS_WINDOWS
+	/* windows ignores the passed name in order to support Unicode */
+	assert(f->f_name != NULL);
+#else
 	assert(name != NULL);
+#endif
 	assert(mode != NULL);
 	assert(f->f_fp == NULL);
 
@@ -156,7 +170,6 @@ open_the_file(PyFileObject *f, char *name, char *mode)
 	else
 #endif
 	{
-		Py_BEGIN_ALLOW_THREADS
 #ifdef WITH_UNIVERSAL_NEWLINES
 		if (strcmp(mode, "U") == 0 || strcmp(mode, "rU") == 0)
 			mode = "rb";
@@ -168,9 +181,27 @@ open_the_file(PyFileObject *f, char *name, char *mode)
 		if (strcmp(mode, "U") == 0 || strcmp(mode, "rU") == 0)
 			mode = "r";
 #endif
+#ifdef MS_WINDOWS
+		if (PyUnicode_Check(f->f_name)) {
+			PyObject *wmode; 
+			wmode = PyUnicode_DecodeASCII(mode, strlen(mode), NULL); 
+			if (f->f_name && wmode) {
+				Py_BEGIN_ALLOW_THREADS
+				/* PyUnicode_AS_UNICODE OK without thread
+				   lock as it is a simple dereference. */
+				f->f_fp = _wfopen(PyUnicode_AS_UNICODE(f->f_name),
+						  PyUnicode_AS_UNICODE(wmode));
+				Py_END_ALLOW_THREADS
+			}
+			Py_XDECREF(wmode);
+		}
+#endif
+		if (NULL == f->f_fp && NULL != name) {
+			Py_BEGIN_ALLOW_THREADS
 			f->f_fp = fopen(name, mode);
 			Py_END_ALLOW_THREADS
 		}
+	}
 	if (f->f_fp == NULL) {
 #ifdef NO_FOPEN_ERRNO
 		/* Metroworks only, wich does not always sets errno */
@@ -201,7 +232,11 @@ open_the_file(PyFileObject *f, char *name, char *mode)
 			PyErr_Format(PyExc_IOError, "invalid mode: %s",
 				     mode);
 		else
+#ifdef MS_WINDOWS
+			PyErr_SetFromErrnoWithFilenameObject(PyExc_IOError, f->f_name);
+#else
 			PyErr_SetFromErrnoWithFilename(PyExc_IOError, name);
+#endif /* MS_WINDOWS */
 		f = NULL;
 	}
 	if (f != NULL)
@@ -215,7 +250,7 @@ PyFile_FromFile(FILE *fp, char *name, char *mode, int (*close)(FILE *))
 	PyFileObject *f = (PyFileObject *)PyFile_Type.tp_new(&PyFile_Type,
 							     NULL, NULL);
 	if (f != NULL) {
-		if (fill_file_fields(f, fp, name, mode, close) == NULL) {
+		if (fill_file_fields(f, fp, name, mode, close, NULL) == NULL) {
 			Py_DECREF(f);
 			f = NULL;
 		}
@@ -293,11 +328,24 @@ file_dealloc(PyFileObject *f)
 static PyObject *
 file_repr(PyFileObject *f)
 {
+	if (PyUnicode_Check(f->f_name)) {
+		PyObject *ret = NULL;
+		PyObject *name;
+		name = PyUnicode_AsUnicodeEscapeString(f->f_name);
+		ret = PyString_FromFormat("<%s file u'%s', mode '%s' at %p>",
+				   f->f_fp == NULL ? "closed" : "open",
+				   PyString_AsString(name),
+				   PyString_AsString(f->f_mode),
+				   f);
+		Py_XDECREF(name);
+		return ret;
+	} else {
 		return PyString_FromFormat("<%s file '%s', mode '%s' at %p>",
 				   f->f_fp == NULL ? "closed" : "open",
 				   PyString_AsString(f->f_name),
 				   PyString_AsString(f->f_mode),
 				   f);
+	}
 }
 
 static PyObject *
@@ -1766,6 +1814,7 @@ file_init(PyObject *self, PyObject *args, PyObject *kwds)
 	char *name = NULL;
 	char *mode = "r";
 	int bufsize = -1;
+	int wideargument = 0;
 
 	assert(PyFile_Check(self));
 	if (foself->f_fp != NULL) {
@@ -1776,12 +1825,33 @@ file_init(PyObject *self, PyObject *args, PyObject *kwds)
 		Py_DECREF(closeresult);
 	}
 
+#ifdef Py_WIN_WIDE_FILENAMES
+	if (GetVersion() < 0x80000000) {    /* On NT, so wide API available */
+		PyObject *po;
+		if (PyArg_ParseTupleAndKeywords(args, kwds, "U|si:file",
+						kwlist, &po, &mode, &bufsize)) {
+			wideargument = 1;
+			if (fill_file_fields(foself, NULL, name, mode,
+					     fclose, po) == NULL)
+				goto Error;
+		} else {
+			/* Drop the argument parsing error as narrow
+			   strings are also valid. */
+			PyErr_Clear();
+		}
+	}
+#endif
+
+	if (!wideargument) {
 		if (!PyArg_ParseTupleAndKeywords(args, kwds, "et|si:file", kwlist,
-					 Py_FileSystemDefaultEncoding, &name,
+						 Py_FileSystemDefaultEncoding,
+						 &name,
 						 &mode, &bufsize))
 			return -1;
-	if (fill_file_fields(foself, NULL, name, mode, fclose) == NULL)
+		if (fill_file_fields(foself, NULL, name, mode,
+				     fclose, NULL) == NULL)
 			goto Error;
+	}
 	if (open_the_file(foself, name, mode) == NULL)
 		goto Error;
 	PyFile_SetBufSize(self, bufsize);

PC/pyconfig.h

@@ -400,6 +400,10 @@ Py_NO_ENABLE_SHARED to find out.  Also support MS_NO_COREDLL for b/w compat */
    Include/unicodeobject.h). */
 #if Py_UNICODE_SIZE == 2
 #define HAVE_USABLE_WCHAR_T
+
+/* Define to indicate that the Python Unicode representation can be passed
+   as-is to Win32 Wide API.  */
+#define Py_WIN_WIDE_FILENAMES
 #endif
 
 /* Use Python's own small-block memory-allocator. */

Python/errors.c

@@ -259,7 +259,7 @@ PyErr_NoMemory(void)
 }
 
 PyObject *
-PyErr_SetFromErrnoWithFilename(PyObject *exc, char *filename)
+PyErr_SetFromErrnoWithFilenameObject(PyObject *exc, PyObject *filenameObject)
 {
 	PyObject *v;
 	char *s;
@@ -314,8 +314,8 @@ PyErr_SetFromErrnoWithFilename(PyObject *exc, char *filename)
 	}
 #endif /* Unix/Windows */
 #endif /* PLAN 9*/
-	if (filename != NULL)
-		v = Py_BuildValue("(iss)", i, s, filename);
+	if (filenameObject != NULL)
+		v = Py_BuildValue("(isO)", i, s, filenameObject);
 	else
 		v = Py_BuildValue("(is)", i, s);
 	if (v != NULL) {
@@ -329,18 +329,40 @@ PyErr_SetFromErrnoWithFilename(PyObject *exc, char *filename)
 }
 
 
+PyObject *
+PyErr_SetFromErrnoWithFilename(PyObject *exc, char *filename)
+{
+	PyObject *name = filename ? PyString_FromString(filename) : NULL;
+	PyObject *result = PyErr_SetFromErrnoWithFilenameObject(exc, name);
+	Py_DECREF(name);
+	return result;
+}
+
+#ifdef Py_WIN_WIDE_FILENAMES
+PyObject *
+PyErr_SetFromErrnoWithUnicodeFilename(PyObject *exc, Py_UNICODE *filename)
+{
+	PyObject *name = filename ? 
+	                 PyUnicode_FromUnicode(filename, wcslen(filename)) : 
+	                 NULL;
+	PyObject *result = PyErr_SetFromErrnoWithFilenameObject(exc, name);
+	Py_XDECREF(name);
+	return result;
+}
+#endif /* Py_WIN_WIDE_FILENAMES */
+
 PyObject *
 PyErr_SetFromErrno(PyObject *exc)
 {
-	return PyErr_SetFromErrnoWithFilename(exc, NULL);
+	return PyErr_SetFromErrnoWithFilenameObject(exc, NULL);
 }
 
 #ifdef MS_WINDOWS 
 /* Windows specific error code handling */
-PyObject *PyErr_SetExcFromWindowsErrWithFilename(
+PyObject *PyErr_SetExcFromWindowsErrWithFilenameObject(
 	PyObject *exc,
 	int ierr,
-	const char *filename)
+	PyObject *filenameObject)
 {
 	int len;
 	char *s;
@@ -362,8 +384,8 @@ PyObject *PyErr_SetExcFromWindowsErrWithFilename(
 	/* remove trailing cr/lf and dots */
 	while (len > 0 && (s[len-1] <= ' ' || s[len-1] == '.'))
 		s[--len] = '\0';
-	if (filename != NULL)
-		v = Py_BuildValue("(iss)", err, s, filename);
+	if (filenameObject != NULL)
+		v = Py_BuildValue("(isO)", err, s, filenameObject);
 	else
 		v = Py_BuildValue("(is)", err, s);
 	if (v != NULL) {
@@ -374,6 +396,36 @@ PyObject *PyErr_SetExcFromWindowsErrWithFilename(
 	return NULL;
 }
 
+PyObject *PyErr_SetExcFromWindowsErrWithFilename(
+	PyObject *exc,
+	int ierr,
+	const char *filename)
+{
+	PyObject *name = filename ? PyString_FromString(filename) : NULL;
+	PyObject *ret = PyErr_SetExcFromWindowsErrWithFilenameObject(exc, 
+	                                                             ierr, 
+	                                                             name);
+	Py_XDECREF(name);
+	return ret;
+}
+
+#ifdef Py_WIN_WIDE_FILENAMES
+PyObject *PyErr_SetExcFromWindowsErrWithUnicodeFilename(
+	PyObject *exc,
+	int ierr,
+	const Py_UNICODE *filename)
+{
+	PyObject *name = filename ? 
+	                 PyUnicode_FromUnicode(filename, wcslen(filename)) : 
+	                 NULL;
+	PyObject *ret = PyErr_SetExcFromWindowsErrWithFilenameObject(exc, 
+	                                                             ierr, 
+	                                                             name);
+	Py_XDECREF(name);
+	return ret;
+}
+#endif /* Py_WIN_WIDE_FILENAMES */
+
 PyObject *PyErr_SetExcFromWindowsErr(PyObject *exc, int ierr)
 {
 	return PyErr_SetExcFromWindowsErrWithFilename(exc, ierr, NULL);
@@ -388,9 +440,29 @@ PyObject *PyErr_SetFromWindowsErrWithFilename(
 	int ierr,
 	const char *filename)
 {
-	return PyErr_SetExcFromWindowsErrWithFilename(PyExc_WindowsError,
-						      ierr, filename);
+	PyObject *name = filename ? PyString_FromString(filename) : NULL;
+	PyObject *result = PyErr_SetExcFromWindowsErrWithFilenameObject(
+						      PyExc_WindowsError,
+						      ierr, name);
+	Py_XDECREF(result);
+	return result;
+}
+
+#ifdef Py_WIN_WIDE_FILENAMES
+PyObject *PyErr_SetFromWindowsErrWithUnicodeFilename(
+	int ierr,
+	const Py_UNICODE *filename)
+{
+	PyObject *name = filename ? 
+	                 PyUnicode_FromUnicode(filename, wcslen(filename)) : 
+	                 NULL;
+	PyObject *result = PyErr_SetExcFromWindowsErrWithFilenameObject(
+						      PyExc_WindowsError,
+						      ierr, name);
+	Py_XDECREF(result);
+	return result;
 }
+#endif /* Py_WIN_WIDE_FILENAMES */
 #endif /* MS_WINDOWS */
 
 void